Document Type



Recommendation systems are the tools whose purpose is to suggest relevant products or services to the customers. In a movie business website, the recommendation system provides users with more options, classify movies under different types to assist in arriving at a decision. Although, with current e-commerce giants focusing on hybrid filtering approach, we have decided to explore the functionality of Content-based recommendation system. This research paper aims to delve deeper into the content-based recommendation system and adding tags to enhance its functionality. The content-based approach is more fit to the movie recommendation as it overcomes the ‘cold start’ issue faced by the collaborative filtering approach, meaning, even with no ratings for a movie, it can still be recommended. The proposed method is to solve the less ‘data categorization’ issue in content-based filtering. Collective Intelligence Social Tagging System (CIST) aims at making a significant difference in content-based recommendation system to enrich the item profile and provide more accurate suggestions. The main gist of CIST is to involve the users to contribute in tagging to build a more robust system in online movie businesses. Tags in the millennial world are the ‘go to’ words that everyone looks up to in an online world of E-commerce. It’s the easiest way of telling a story without actual long sentences. We recommended three main solutions for the concerns of CIST, (a) clustering of tags to avoid synonymous tag confusion and create a metadata for movies under same tags, (b) 5 criteria model to motivate and give the most amount of genuine information for end users to trust and eventually contribute in tagging, and (c) clear way of distinguishing and displaying tags to separate primary tags and secondary tags and give a chance to the users to assess whether the given tags reflect the relevant theme of the film.